Guest Lecture for PPS 3250

The Effect of Same-Sex Marriage Legalization on Adoptions & Family Formation

Joshua C. Martin
joshua.martin@vanderbilt.edu

Vanderbilt University

February 24, 2025

Identifying a research topic

Pay Attention to What You Pay Attention To

Identifying a research topic

Pay Attention to What You Pay Attention To

Identifying a research topic

Pay Attention to What You Pay Attention To

Identifying a research topic

Pay Attention to What You Pay Attention To

Identifying a research topic

Pay Attention to What You Pay Attention To

Identifying a research topic

Pay Attention to What You Pay Attention To

Identifying a research topic

Pay Attention to What You Pay Attention To

Identifying a research topic

Pay Attention to What You Pay Attention To

Identifying a research topic

Find Inspiration

Identifying a research topic

Find Inspiration

Identifying a research topic

Theorizing the relationship

Identifying a research topic

Theorizing the relationship

Data

Finding the data

Data

Cleaning the data

  • 25m \(\times\) 200 data frame

  • What level should the analysis ideally be preformed at?

  • What level can the analysis be preformed at?

Data

Creating a panel dataset

  • Assign a date for everyone

    • Date of discharge
    • Current date
    • Latest date
    • First removed

  • Time in foster care = Date – first removed

Data

Creating a panel dataset

  • I chose to aggregate the data

    1. Computational simplicity

    2. Controlling for the population of children in foster care

Data

Dependent variables

  • Data frame creation (County by State by Year)

  • Tabulation = number of children interacting with the foster care system per year

    • Total Discharges = Aggregate number of individuals with any “discharge reason”
    • Parental Reunifications
    • Adoptions = living with Other relatives, adoption or guardianship
    • “Bad” Discharges = emancipation, transfer to Another Agency, runaway, or death
    • Still in system = foster care – total discharges

Early Work

Research is a trial and error process

  • I bit off more than I could chew – simultaneously studing the impact of:

    • Same-sex marriage, Domestic partnership / civil unions, Legislative Bans, Constitutional Bans

  • Extrapolated evidence that I did not provide

  • I introduced a “novel” new empirical method

    • 25% of the paper was methods

Early Work

Research is a trial and error process

Early Work

Research is a trial and error process

Early Work

Research is a trial and error process

Data

Plot it out

Analysis

Plot it out

Analysis

Plot it out

In-Class Exercise

Massachusetts & Vermont

  • What is the appropriate “time window”

    • All states eventually legalize SSM

  • Do we need to omit any “treated” states from this time window?

  • Create an indicator variable for “treatment” (SSM)

  • What dependent variable should we use?

  • Do we need any control variables